Non-contacting face seals are usually applied to high-speed high-pressure rotating equipment, where the use of ordinary mechanical face seals with face contact would result in excessive generation of heat and wear. Non-contacting operation avoids this undesirable face contact when the shaft is rotating above a certain minimum speed, which is called a lift-off speed.
As with ordinary mechanical seals a non-contacting face seal consists of two sealing rings, each of which is provided with a very precisely finished sealing surface. These surfaces are perpendicular to and concentric with the axis of rotation. Both rings are positioned adjacent to each other with the sealing surfaces in contact at conditions of zero pressure differential and zero speed of rotation. One of the rings is normally fixed to the rotatable shaft, the other located within the seal housing structure and allowed to move axially. The seal housing structure is normally stationary and fixed to the body of the compressor, turbine or pump to which the seal is applied.
To achieve non-contacting operation of the seal, one of the two sealing surfaces in contact is normally provided with shallow surface recesses, which act to generate pressure fields that force two sealing surfaces apart. When the magnitude of the forces resulting from these pressure fields is large enough to overcome the forces that hold seal faces closed, the sealing surface will separate and form a clearance, resulting in non-contacting operation. The character of the separation forces is such that their magnitude decreases with the increase of face separation. Opposing or closing forces, on the other hand, depend on sealed pressure level and as such are independent of face separation. They result from the sealed pressure and the spring force acting on the back surface of the axially movable sealing ring. Since the separation or opening force depends on the separation distance between sealing surfaces, unlike the closing force, during the operation of the seal or on imposition of sufficient pressure differential, equilibrium separation between both surfaces will establish itself. This occurs when closing and opening forces are in equilibrium and equal to each other.
Equilibrium separation constantly changes within the range of gaps. The goal is to have the low limit of this range above zero. Another goal is to make this range as narrow as possible, because on its high end the separation between the faces will lead to increased seal leakage.
Since non-contacting seals operate by definition with a clearance between sealing surfaces, their leakage will be higher then that of a contacting seal of similar geometry. Yet, the absence of contact will mean zero wear on the sealing surfaces and therefore a relatively low amount of heat generated between them. It is this low generated heat and lack of wear that enables the application of non-contacting seals to high-speed turbomachinery, where sealed fluid is gas. Turbocompressors are used to compress this fluid and since gas has a relatively low mass, they normally operate at very high speeds and with a number of compression stages in series. This requires relatively long shafts that have to operate frequently above their critical speeds. Neither seals nor bearings can contact at these high speeds so shafts are therefore made to float on fluid cushions, both at the bearings and the seals.
In the above situation, it becomes important to make sure that the shaft does not vibrate excessively during operating speeds as well as during acceleration or deceleration. This means that the shaft inertia axis must be as close to its geometrical axis as possible, since the bearings will be forcing shaft rotation around its axis of geometry rather than around its axis of inertia.
Should there be a significant distance between the inertia and geometry axes, the shaft would be prone to vibration, making it more likely to contact stationary parts in its vicinity. Such contact may result in overheating, wear and perhaps a failure at the bearing or the seal. To prevent the above problem, shafts are normally balanced by a selective removal of shaft material or by a redistribution of weights at its periphery. This balancing thus eliminates the effects of machining inaccuracy and/or material non-uniformity.
Since the shaft rotates with components attached to it, such as compressor impellers, bearing sleeves and seal rings, it is necessary to balance it with these components in place. While many can be fixed to the shaft during its manufacture and therefore balanced in the manufacturer's shop, rotating parts of the non-contacting seal normally come as a part of the seal cartridge and are mounted onto the shaft on site in the process plant and thus away from the manufacturer's shop. Special care, therefore, has to be exercised to keep the unbalance of the rotating seal parts to a minimum. A key requirement in this regard is that rotating parts remain concentric to the shaft at all times.
This is not easy to accomplish, because seal components are often of diverse materials with widely different coefficients of thermal expansion and moduli of elasticity. As the machine goes through thermal and pressure cycles, clearances between concentric components may vary.
One way of maintaining the concentricity of components despite variation in clearance is to place flexible elements between both concentric components. As the clearance changes, the flexible element deforms elastically an equal amount all around and thus maintains concentricity.
One example of such a solution is prior art to U.S. Pat. No. 5,039,113. This patent shows how a rotating sealing ring can be supported by an elastic strip at its inner diameter. It also illustrates how two such elastic strips can be placed close to both ends of a relatively long sleeve, between its inner diameter and the shaft.